1 Introduction

The intrinsic passivation layers play a key role to reduce the surface state density of the silicon wafer and improve the photoelectric conversion efficiency of the crystalline silicon heterojunction solar cells [1]. It is one of core technologies behind the record conversion efficiency for HIT solar cell of Panasonic Corporation [2]. In which oxygen doped amorphous silicon (a-SiOx:H) is a good choice for the material of the intrinsic thin-layer [1]. It is well known that the substrate temperature during the film deposition has great effect on the properties of the a-SiOx:H films [3], and on the effect of the film on the c-Si wafer passivation [4]. Muller et al. [5] also showed that the low temperature (250 °C) post-annealing process has effect on the passivation of a-SiOx:H film on c-Si wafer. For the mechanism of the temperature of the thermal processing affecting on the a-SiOx:H passivating the c-Si wafer surface, there is no exact explanation. Jaran et al. [6] mentioned that the Si–(OH)x, Si–O–Si bonds in the film have some effect to reduce the defects in the films and improve the passivation effect. It is thought that this is not exact enough to show the mechanism of the c-Si surface passivation by a-SiOx:H, for the Si–Hx bonds [7], crystalline quality [8] also changed during the thermal processing. In order to understand the performance and the mechanism more clearly, some series of samples with widely arranged temperatures for the as-deposition and post-annealing processes for the a-SiOx:H films on c-Si wafer were designed. The samples were analyzed with Spectroscopic Ellipsometry (SE) and Fourier Transform Infrared Spectroscopy (FTIR), in order to understand the influence the temperature of deposition and post-deposition annealing have on the property and performance of the a-SiOx:H films more unequivocal.

2 Experimental

The silicon wafers used are n type Czochralski monocrystalline Si (100), with a thickness of 180 μm, resistivity of 3 Ω cm and a size of 40 × 40 mm2. Before passivation layers deposition, the wafers were cleaned by acetone, then etched about 25 μm each side by KOH solution (20 wt%, 85 °C) for 20 min, then cleaned by SC1 solution (80 ml H2O2 + 80 ml NH4OH + 400 ml H2O, 80 °C) for 10 min and SC2 solution (80 ml H2O2 + 80 ml HCl + 400 ml H2O, 80 °C) for 10 min, last dipped in HF solution (2 vol%) for 1 min to removing the native oxide layers.

The passivation layer is hydrogenated amorphous silicon sub-oxides (a-SiOx:H) on both sides of the n-Cz-Si wafers by AC glow discharge of plasma (13.56 MHz, CCP type) enhanced chemical vapor deposition (PECVD, HW-PECVD-1E, Weina Ltd. China), and the source gases are silane (SiH4), hydrogen (H2) and carbon dioxide (CO2). All the thicknesses of the deposited layers are about 100 nm. Gas flux ratios are SiH4:H2:CO2 = 3:15:0.6 (sccm), air pressure is 22 Pa; power is 12 W, 10 min.

After deposition, some samples are post-annealed for 1 h by Halogen-lamp heating furnace, in argon atmosphere and under atmospheric pressure.

The lifetime of the samples before and after the post-annealed processes was tested by RF-PCD method, Sinton WCT-120. The hydrogen bonds structure of the passivation layers was tested by Fourier infrared spectrometer (FTIR, Nicolet 380). The imaginary dielectric factor of the wafers was analyzed by Spectroscopic Ellipsometry (Semilab, GES5-E).

3 Results and discussions

3.1 The influence of the post-deposition annealing temperature

The post-deposition annealing temperatures of each wafer are 200, 250, 275, 300, and 350 °C respectively. The effective lifetime of an as-deposited wafer and all the post-annealed wafers are shown in Fig. 1.

Fig. 1
figure 1

The effective lifetime of n-Cz-Si wafers passivated by a-SiOx:H. a As-deposited; b post-annealed at different temperatures

Full size image

Figure 1 shows the results of effective lifetime of n-Cz-Si wafers passivated by a-SiOx:H, with and without post-annealing. Figure 1a shows that passivation effect of a-SiOx:H deposited at room temperature is poor. The post-deposition annealing has great effect on improving the passivation effect of the a-SiOx:H layers as Fig. 1b, for the lifetime of the post-annealed wafer increases greatly. The maximum value of the lifetime of the wafers belongs to the wafer post-annealed at 275 °C, the value is 1371 μs (at a minority carrier density of 1 × 1015 cm−3). For calculation of surface recombination velocity (SRV) S eff, it is assumed that both surfaces provide a sufficiently low recombination velocity (less than D/4w [9, 10], D is the diffusion coefficient, about 35 cm2/s, w is the thickness of wafer, about 180 μm), and have the same values S eff = S front = S back, as the sample structure is symmetric. The surface recombination velocity can be deduced by Eq. (1) [5]:

$$S_{eff} = \left( {\frac{1}{{\tau_{eff} }} - \frac{1}{{\tau_{bulk} }}} \right) \times \frac{w}{2}$$
(1)

In which w is the wafer thickness, τ bulk is set infinity (an upper limit of the SRV is calculated for the case that no SRH recombination is considered by setting τ bulk  → ∞ in Eq. 1). From Eq. (1) Seff of the wafer annealed at 275 °C is 4.7 cm/s. The results of imaginary part of dielectric constant (ε 2) and Raman spectra of samples are shown in Fig. 2.

Fig. 2
figure 2

a The ε 2 of the annealed wafers with different annealing temperature; b Raman spectra

Full size image

Figure 2a shows ε 2 spectra of n-Cz-Si passivated by a-SiOx:H at room temperature without and with annealing. The spectra of the samples with annealing temperature below 300 °C show peaks at about 3.6 eV, are typical of amorphous silicon characteristic spectrum [8, 11, 12]. The Raman results of Fig. 2b show that a-Si peak of the films covered wafers appear obviously around 480 cm−1 compare to that of the naked Si wafer, but the peak density of the Si–O–Si bond does not depend on a-SiOx:H film. It is unambiguous that the films in this series are amorphous silicon structure without detectable SiO2 structure [5, 10]. With the increasing of the thermal annealing temperature, ε 2 peak slightly shifts to higher energy, but significantly turns to lower energy when the temperature is higher than 300 °C. It can be explained that the SiHn stretching modes change in the films [13, 14], the FTIR spectra and microstructure parameter R *of the samples are fitted by “multiple peak fit” function of Origin in Fig. 3. The R * defined by Eq. (2) [15]

Fig. 3
figure 3

a SiHn of the samples with different post-annealing temperatures; b microstructure parameters R *

Full size image
$$R^{*} = \frac{{I_{2090} }}{{I_{2000} + I_{2090} }}$$
(2)

Figure 3 is SiHn of the samples with different post-deposition annealing temperatures and microstructure parameters R *. The ratio of the fitted peaks area can be characterized with the relative content of SiH and SiH2 bonds in the a-SiOx:H films [2]. The peaks are at 2000 and 2090 cm−1 corresponds to SiH bond and SiH2 bond, respectively. Because of oxygen attachment, the peak of SiH2 shifts to a higher wave number [6]. Figure 3 can infer that the passivation effect is the best when R * at an appropriate proportion.

3.2 The co-operation effect of the substrate-heating during the deposition process and the post-deposition annealing process

This series of samples is deposited at different substrate heating temperatures, and post-annealed at 275 °C for 1 h. The substrate heating temperature during film deposition are 25, 100, 180, 200, 220, and 250 °C, respectively. The tested effective lifetimes of the wafers as-deposited and post-annealed are shown in Fig. 4.

Fig. 4
figure 4

a The lifetime of the same wafers with a-SiOx:H layers as-deposited at different substrate temperatures; b annealed at 275 °C

Full size image

Figure 4 shows the effective lifetime of the post-annealed n-Cz-Si wafers with passivation layers at 275 °C. The insert shows the lifetime of the same wafers with a-SiOx:H layers as-deposited at different substrate temperatures. It is clear that with the substrate temperature increasing, the effective lifetime increases firstly and then decreases, the optimum passivation effect is gained at the substrate temperature of 220 °C with the highest lifetime of 747 μs (at a minority carrier density of 1 × 1015 cm−3). The result is similar with the optimum temperature 210 °C in Ref. [9].The post annealing process at 275 °C has great effect on improving the effective lifetime of most of the samples significantly. The highest effective lifetime value is 2135 µs (at a minority carrier density of 1 × 1015 cm−3) and belongs to the sample with the films deposited at 100 °C. Since the thickness of the wafer is 130 μm, the surface recombination velocity is 3.1 cm/s according to the Eq. (1), which is good enough for silicon heterojunction solar cells. The ε 2 and Raman spectra of samples are shown in Fig. 5.

Fig. 5
figure 5

a The ε 2 spectra of samples before and after the post-annealing process; b the Raman spectra of samples after post-annealing

Full size image

Figure 5a is the ε 2 spectra of n-Cz-Si wafers before and after the post-annealing deposited a-SiOx:H films with different temperature. All the spectra show a broad peak at around 3.6 eV, which shows that thin films in the samples, are amorphous silicon structured [8, 16]. The hollow symbols in Fig. 6 represent the ε 2 spectra of wafers after annealing with 275 °C. It is shown that the amplitude and peak position decrease with increasing post-annealing temperature, significantly for the samples of deposited temperature higher than 200 °C. It can be explained that with the increasing of temperature, these trends are consistent with hydrogen in the films reducing at higher temperatures [15]. From the Raman spectra of the samples shown in Fig. 5b, it can be further clarified that the films on the samples are mainly amorphous silicon structure. The SiHn microstructure and its parameter R *of samples are shown in Fig. 6.

Fig. 6
figure 6

a SiHn bonds of the samples with different deposition temperatures and the same post-annealing 275 °C; b microstructure parameters R*

Full size image

Figure 6 indicate that SiHn bonds composition and microstructure parameter R* change with deposition temperature increasing, although the samples underwent the same post-annealing process. According to Figs. 3 and 6, passivation effect is optimum when R * is about 0.67, the performance of a-SiOx:H films is highly sensitive to the content of H and SiHn bonds composition [13, 17], the H content and SiHn bonds composition of a-SiOx:H film is closely affected by the thermal processes, both deposition and post-annealing.

4 Conclusions

In order to understand the effect of temperature of the deposition and post-annealing processes on intrinsic a-SiOx:H film for n-Cz-Si wafer passivation, three series samples with different film deposition temperature or/and different post-annealing temperature are prepared and analyzed. It can be concluded that: (1) the structure and passivation effect of a-SiOx:H films on n-Cz-Si wafer are sensitive to the temperature of the deposition and post-annealing processes, the optimum scheme is depositing at 100 °C and post-annealing at 275 °C; (2) the microstructure parameter R *of the a-SiOx:H is ~0.67 for the samples with the optimum passivation effect.